The thesis
Bollinger-reversion is a classic mean-reversion play: buy when price drops below the lower Bollinger band, sell when it pushes above the upper band. It trades a universe of 24 US large caps — the usual megacap tech (AAPL, MSFT, GOOGL, NVDA), financials (JPM, BAC, V, MA), healthcare, staples and energy. The bet is that short-term dislocations in liquid names snap back toward their moving average. It's a well-trodden idea, which is both its appeal and its risk: everyone knows it, so any edge is thin.
Recent activity
The strategy is live, but lately it has been quiet. Across the last six scheduled runs (Nov 3–10), most days produced zero or one execution, with one rejection on Nov 9. Recent fills include buys in XOM (15 @ $166.26) and NVDA (13 @ $184.15), a round-trip in AAPL (bought 7 @ $329.82, sold 7 @ $352.27 — a clean reversion win), and a trimmed MSFT position. As of the Nov 10 run the book stood at $9,743.54 total with $1,055.32 in cash — modestly below its roughly $10k starting point, so live results are flat-to-slightly-negative rather than tracking the backtest's gains.
Backtest performance
Over 451 days the backtest returned +17.55% (final equity $11,755), a 9.46% CAGR, with a 63.89% win rate across 76 trades. That hit rate is the strongest selling point. But the supporting stats are soft: a Sharpe of just 0.66 and a 20.57% max drawdown mean the ride was volatile for the reward. Turnover of 1,721% is heavy churn, though fees stayed low at $76 and FX cost was zero.
Validation — the red flag
This is where the story turns. Walk-forward validation failed. All four folds were positive, which sounds reassuring until you read the trend:
| Fold | Return | Sharpe |
|---|---|---|
| 1 | 7.67% | 1.78 |
| 2 | 1.42% | 0.25 |
| 3 | 0.77% | 0.20 |
| 4 | 0.40% | 0.14 |
The edge decays monotonically — nearly all the profit came from the earliest window, and the most recent fold produced a 0.40% return at a 0.14 Sharpe. Out-of-sample stats confirm it: OOS return of 0.4% and OOS Sharpe of 0.14. The deflated Sharpe ratio (DSR 0.342) sits well below the threshold where we'd call the result statistically real after accounting for the 6 trials run. PSR of 0.814 looks healthier, but DSR is the number that penalizes selection bias, and it isn't convinced.
Verdict
Strengths: an intuitive, cheap-to-run thesis and a genuinely high win rate. Risks: the headline return is almost entirely a historical artifact of one favourable window, out-of-sample performance is indistinguishable from flat, and live trading so far echoes that — the book is treading water. The honest read is that bollinger-reversion looks overfit. It stays live as a paper experiment worth watching, but on current evidence it has not earned a capital allocation.